VoidSakura Pulse

A self-hosted metrics and alerting platform for game and application servers, with per-tick profiling and an alerting model that respects on-call hours.

Observability Infrastructure Performance

Current version: 1.8.02 Aug 2026
Download: 8

$39.00

One-time purchase. Updates included.

Documentation

Signed, expiring downloads
Every future update included
Support from the people who built it

Pulse collects, stores and alerts on the signals that actually predict an incident: tick time distribution, garbage collection pauses, connection churn, queue depth and disk latency.

It ships as a single binary with an embedded time-series store. There is no Prometheus to operate, no Grafana to configure and no agent fleet to keep in sync — one process, one configuration file, one dashboard.

Alerting is built around sustained conditions rather than instantaneous thresholds, and every alert names the metric, the window and the value that triggered it. Routing supports quiet hours and escalation so a 3 a.m. page has to earn itself.

What it does

Embedded time-series store

A compressed columnar store with configurable retention. No external database, no separate scrape infrastructure.

Tick-level profiling

Sub-tick sampling for game servers, attributing time to plugins, entities and chunk work rather than reporting one aggregate number.

Sustained-condition alerts

Alerts fire on a condition holding across a window, and each notification states the metric, window and observed value.

Quiet hours and escalation

Routing rules understand on-call schedules, so low-severity conditions wait for morning instead of waking someone.

Single binary deployment

One process with one configuration file. systemd unit and Docker image included, both with health and readiness endpoints.